Discussion:
[SlimDevices: Audiophiles] Alternatives to MQA - "Bit freezing"
Julf
2016-09-04 19:42:25 UTC
Permalink
Jim Lesurf, "Science Journalist" for Hi-Fi News, has proposed a new,
simplified method of accomplishing pretty much what MQA accomplishes, by
using a method he calls "bit freezing" - basically truncating the lowest
7 bits off 24-bit material, as they only contain random noise anyway.
That random noise doesn't contribute anything to the actual music, but
makes it much harder for FLAC or ALAC to compress the files. Dropping
the unnecessary random bits makes the size of the files up to 75%
smaller, accomplishing the same result as MQA without the need for any
resampling.

Yes, we might all go "duh, yeah!", but here is at least one journalist
from a mainstream hi-fi magazine admitting the extra 8 bits in 24-bit
material is just random noise (an assumption MQA also makes).

link: 'bitfreezing.html'
(http://www.audiomisc.co.uk/MQA/cool/bitfreezing.html)



"To try to judge the real from the false will always be hard. In this
fast-growing art of 'high fidelity' the quackery will bear a solid gilt
edge that will fool many people" - Paul W Klipsch, 1953
------------------------------------------------------------------------
Julf's Profile: http://forums.slimdevices.com/member.php?userid=42050
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
Wombat
2016-09-05 00:58:35 UTC
Permalink
Use Lossywav and you'll see how low the data becomes you really need but
this makes sense.
MQA is this magic De-Blurrring you need to surpass everything you ever
knew and bitfreezing doesn't.



Transporter (modded) -> RG142 -> Avantgarde Acoustic based 500VA
monoblocks -> Sommer SPK240 -> self-made speakers
------------------------------------------------------------------------
Wombat's Profile: http://forums.slimdevices.com/member.php?userid=4113
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
drmatt
2016-09-05 12:47:00 UTC
Permalink
I can see the point in NOT saving tens of bits of random entropy in my
audio files, but isn't bit freezing just a funky name for truncation...




------------------------------------------------------------------------
drmatt's Profile: http://forums.slimdevices.com/member.php?userid=59498
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
Julf
2016-09-05 13:03:21 UTC
Permalink
Post by drmatt
I can see the point in NOT saving tens of bits of random entropy in my
audio files, but isn't bit freezing just a funky name for truncation...
That is what it looks like to me...



"To try to judge the real from the false will always be hard. In this
fast-growing art of 'high fidelity' the quackery will bear a solid gilt
edge that will fool many people" - Paul W Klipsch, 1953
------------------------------------------------------------------------
Julf's Profile: http://forums.slimdevices.com/member.php?userid=42050
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
Archimago
2016-09-06 06:27:51 UTC
Permalink
Interesting but not surprising.

Yeah, truncate down to 17-bits... Maybe do a little dithering in that
last LSB. Keep the samplerate at 96kHz. Then implement a gentle filter
above 25kHz or so... Voila, higher than CD dynamic range. Good
compression ratio. Probably as good as MQA.

No surprise of course. And we can dream up of all kinds of variations to
these schemes. The only question is whether MQA has any "secret sauce"
in their DSP processing and this whole "de-blur" process as per Wombat.

After close to 2 years, I'm mainly interested in 2 questions: "Where's
MQA?" And "Is it DOA?"



Archimago's Musings: (archimago.blogspot.com) A 'more objective'
audiophile blog.
------------------------------------------------------------------------
Archimago's Profile: http://forums.slimdevices.com/member.php?userid=2207
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
Julf
2016-09-07 06:15:20 UTC
Permalink
"Bit freezing" seems like FUD , just do a proper resampling with dither
to 16-18 bits There about and pick some sample rate slightly higher than
44.1 kHz and be done.
I asked Jim Lesurf about that on uk.rec.audio, and he admitted that that
is precisely what his "bit freezing" is.



"To try to judge the real from the false will always be hard. In this
fast-growing art of 'high fidelity' the quackery will bear a solid gilt
edge that will fool many people" - Paul W Klipsch, 1953
------------------------------------------------------------------------
Julf's Profile: http://forums.slimdevices.com/member.php?userid=42050
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
Julf
2016-09-07 10:28:13 UTC
Permalink
The only "advantage" if you can call it that is that a bit freeze allows
you to stick with a standardised packaging method such as 24/48k flac
and still get the space compression benefit. So some recordings that are
worthy of it would retain 20 "unfrozen" bits, the rest would retain only
16 or some even less.. flac will handle the easy job of deflating all
those zeroes so they don't really exist. Perhaps someone can write a
flac processor that can determine how many bits if signal need to be
retained in a file and simply truncate and reencodes to save space..?
Wouldn't that require AI to determine what is random noise and what is
music? Not that I have come across any commercial recordings worthy of
20 bits...



"To try to judge the real from the false will always be hard. In this
fast-growing art of 'high fidelity' the quackery will bear a solid gilt
edge that will fool many people" - Paul W Klipsch, 1953
------------------------------------------------------------------------
Julf's Profile: http://forums.slimdevices.com/member.php?userid=42050
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
Mnyb
2016-09-07 16:26:01 UTC
Permalink
Ues thats a point a pop or rock song can be bundle of adc and shrug
sampling
:) deblur that



--------------------------------------------------------------------
Main hifi: Touch + CIA PS +MeridianG68J MeridianHD621 MeridianG98DH 2 x
MeridianDSP5200 MeridianDSP5200HC 2 xMeridianDSP3100 +Rel Stadium 3
sub.
Bedroom/Office: Boom
Kitchen: Touch + powered Fostex PM0.4
Misc use: Radio (with battery)
iPad1 with iPengHD & SqueezePad
(spares Touch, SB3, reciever ,controller )
server HP proliant micro server N36L with ClearOS Linux

http://people.xiph.org/~xiphmont/demo/neil-young.html
------------------------------------------------------------------------
Mnyb's Profile: http://forums.slimdevices.com/member.php?userid=4143
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
Wombat
2016-09-07 16:38:28 UTC
Permalink
I see the same pattern as before with people that reported to hear 24bit
advantages in old Jazz music that was pulled from sale for being padded
16bit. Later they heard advantages in dsd and now welcome MQA but at the
same time are completely overwhelmed with the most simple abx because of
to many knobs.
What puzzles me even more is that lately some audiophile claimed
recordings to me sound simply mudded. Means like it purposely does mask
anything that may sound offensive even when the instrumentation in
reality needs to. Perfect digital recordings get EQ'd to death or make a
round to tape and back. I may hear extremely small details on living
room recordings done on a simple digital mixer from a bandcamp newcomer
and miss the kick at my latest 24/96 purchase coming from a so called
audiophile label.



Transporter (modded) -> RG142 -> Avantgarde Acoustic based 500VA
monoblocks -> Sommer SPK240 -> self-made speakers
------------------------------------------------------------------------
Wombat's Profile: http://forums.slimdevices.com/member.php?userid=4113
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
arnyk
2016-09-07 20:57:59 UTC
Permalink
Some kind of entropy analysis , i think most recordingsbwould need less
than 16 bits :/
Before word length mania was even a sparkle in the usual suspect's eye,
(ca. 1975-1980) BBC studied the matter and set their limit at 13 bits.

The source material was professionally originated analog material low
pass filtered at 15 KHz. The effect that was being addressed may be
related to the difference between standard practice for analog and
digital filters of these kinds.

Samples prepared from high sample rate recordings around 2003 but
downsampled to far lower frequencies can be found here:

http://web.archive.org/web/20041204101025/http://pcabx.com/technical/low_pass/castanets-060_lo22Khz.wav

and here:

http://web.archive.org/web/20041210034445/http://www.pcabx.com/technical/low_pass/index.htm


------------------------------------------------------------------------
arnyk's Profile: http://forums.slimdevices.com/member.php?userid=64365
View this thread: http://forums.slimdevices.com/showthread.php?t=106085
Loading...